Enterprise Portfolio

A New Revenue Category for Nera

Nera Telecom has historically sold infrastructure: connectivity, hardware, and managed networks. While this business is stable, margins are compressing and the growth ceiling is visible. Transitioning to enterprise AI services for Indonesia's largest state-owned enterprises (BUMN) represents a new, high-margin revenue category.

Three accounts are now in active engagement. Each one has a specific operational problem that existing tools do not solve. Ancol cannot see its 552-hectare park in real time. AirNav's engineers search filing cabinets for procedures during time-critical situations. Kementan's ministry leadership has no reliable way to track budget execution across 34 provinces.

To deliver on these opportunities, Bireka proposes the NayaAI platform, a production running AI system running across multiple industries, including palm oil plantations, industrial mills, and environmental monitoring systems since 2024. This proposal asks the board to fund three PoC-to-MVP engagements that convert these open doors into recurring service contracts.

How the Partnership Works

Nera is the prime contractor. Nera owns the client relationship, handles commercial billing, provides infrastructure (edge servers, Teltonika IoT SIMs, network links), and manages the SLA. Bireka builds and operates the AI. KMJ, an existing Nera partner, holds the Ancol account and provides field coordination for AirNav.

Every AI deployment pulls Nera hardware revenue behind it: each site requires an edge server, IoT SIMs, and secure network links. The software is the wedge. The infrastructure is the annuity.

NayaAI Is Not a Dashboard

Most enterprise AI vendors sell dashboards. The client gets a screen full of charts and is expected to figure out what they mean. NayaAI works differently. It is an AI brain that sits inside the client's facility, watches multiple data sources simultaneously, and pushes its conclusions directly to the people who need them, through WhatsApp.

No apps to install. No staff to retrain. No data sent to the cloud. The AI talks to managers the same way their teams already do: in Bahasa Indonesia, via their existing messaging apps (WhatsApp, Telegram), with the context of what is actually happening on the ground right now.

The diagram below shows how this works for Ancol. The same architecture applies to every client, with different "senses" depending on the use case.

The data sources are infrastructure. The AI brain is the product. A camera is just a camera until the AI watches it. A chat group is just noise until the AI reads it, classifies every message, and cross-references what it reads with what it sees on camera.

Three Things Only the AI Can Do

Situational fusion. No human can watch 50 cameras, read 500 WhatsApp messages, and track 20 gates at the same time. The AI does this continuously, and produces a single-paragraph situational summary on demand.

Prediction. The AI combines real-time flow data with weather forecasts, historical patterns, and ticket pre-sales to predict what will happen 15, 30, and 60 minutes from now. "Parking C will be full in 18 minutes." "Rain at 3pm, expect mass exit from outdoor zones."

Proactive recommendations. The AI does not wait to be asked. When it sees something forming that nobody has noticed yet, it pushes an alert to the relevant manager with a specific recommended action. This is the moment that sells the product.

552 Hectares, Zero Visibility

PT Pembangunan Jaya Ancol operates the largest integrated tourism destination in Southeast Asia. On peak days, 77,000 visitors fill a 552-hectare site that includes Dufan theme park, beaches, hotels, a marine park, and an eco-park. Revenue in the first half of 2025 reached Rp 495 billion.

The problem is not a lack of data. Ancol already has CCTV cameras across the park, gate counters at every entrance, a fleet of shuttles and service vehicles, and WhatsApp groups where parking staff, security, F&B, and transport teams report conditions throughout the day. The problem is that nobody can process all of it at once. By the time a manager reads a WhatsApp message about a full parking lot, the lot has been full for 20 minutes and 300 visitors have been circling.

NayaAI connects to these four existing data sources and fuses them into a single operational picture. It watches 20 cameras using scoped computer vision (analyzing queue frames every 3 minutes), reads 5 operations chat groups, ingests gate counter data from 5 entries, and tracks 10 vehicles through Teltonika GPS. Then it talks to managers through their preferred enterprise messaging app, in the language and medium they already use.

Operations in One View

Here is what that looks like in practice:

The AI read three messages from three different WhatsApp groups, classified each one by type and urgency, cross-referenced the queue overflow report with live camera data to confirm the crowd estimate, and presented the whole picture as a structured briefing. No human was monitoring those groups.

Proactive Intelligence

The conversation on the previous page shows the AI responding to a question. What follows is more important: the AI noticing something nobody asked about and pushing an alert on its own.

The AI noticed a discrepancy: cameras showed crowd density rising sharply, but gate counters showed no spike in new entries. It concluded that visitors were migrating between zones (likely for the 3pm show) and predicted the food court would be overwhelmed in 12 minutes. No human would catch that pattern across 20 cameras and 5 gate feeds simultaneously.

This is the demo moment. When the ops manager receives a prediction like this and it turns out to be right, the product sells itself.

Execution & Resources

Ancol Delivery Plan

The PoC runs on one zone (Dufan): 5 cameras, 2 WhatsApp groups, 2 gates, 3 GPS vehicles, and 3 AI users. The PoC success criterion is simple: the AI predicts a parking or congestion event before staff report it. The MVP expands to all 5 zones, 20 cameras, 5 groups, and 10 vehicles.

Manpower Per Phase

Resources Required

Commercial Summary

The total execution cost for PT Pembangunan Jaya Ancol is consolidated below. It includes all manpower, site assessment, discovery workshops, custom models development, custom dashboard integration, hardware delivery, license provision, and client hand-off training.

847 Assets, Filing Cabinet Search

AirNav Indonesia (Perum LPPNPI) manages air navigation services for the entire Indonesian archipelago: 296 airports, 847 pieces of critical infrastructure (radar, VOR, ILS, communication systems), and over 1.9 million flights per year. Their engineering teams maintain this equipment under strict safety regulations, with thousands of SOPs, NOTAMs, manuals, and maintenance records accumulated over decades.

The situation on the ground: an engineer who needs to look up a procedure opens a filing cabinet or searches scattered PDF folders. Equipment health is tracked through spreadsheets and manual logbooks. When a piece of radar equipment shows a pattern of escalating alarms, that pattern is invisible until it causes an outage.

NayaAI delivers two AI systems for AirNav, both deployed on the client's own servers, with zero connection to any live ATC, radar, communication, or navigation system. The AI is read-only, physically isolated, and advisory only.

Pattern Recognition

The engineer asked a casual question. The AI searched 24 months of alarm history, detected an escalation pattern, matched it against a known failure event at a different site, calculated a health score, and recommended a specific inspection focus. With source citations. In under a second.

Knowledge Copilot

In time-critical air traffic control environments, supervisors and engineers must reference complex documentation under severe pressure. The Knowledge Copilot indexes all SOPs, NOTAMs, manuals, and circulars to answer procedural queries instantly, with complete source transparency.

Instead of searching through binders, the supervisor got a step-by-step procedure with the exact SOP reference, a historical precedent, and a confidence score. The Knowledge Copilot indexes 1,500+ documents (SOPs, NOTAMs, manuals, circulars) for the MVP.

AirNav Delivery & Pricing

The PoC covers 1 site (JATSC Jakarta): 350 documents ingested, 12 months of alarm data, 8 AI users across operations, engineering, and management. The MVP expands to 3 sites, 1,500+ documents, 24 months of data, and 18 users including executive access.

Manpower Per Phase

Resources Required

Commercial Summary

Execution Cost

The total execution cost for AirNav Indonesia is consolidated below. It includes the complete edge deployment on-premise, security hardening, penetration testing, and bilingual document processing pipelines.

The Largest Opportunity, Done Right

Kementan is a national ministry with operations across 34 provinces and a mandate for digital transformation. It is the largest opportunity of the three. It is also the only one where we do not yet know what the AI would work with.

Unlike Ancol (where we can see the cameras and read the WhatsApp groups) or AirNav (where alarm data and SOPs exist in identifiable systems), Kementan's internal data landscape is uncharted. We do not know what systems exist, what format the data is in, or whether it is complete enough for AI to use. Committing a Rp 500M PoC budget to an unverified environment is not responsible.

Instead, we propose a Rp 239M Strategic Blueprinting that maps the data, identifies where NayaAI fits without competing with Kementan's own TANIA system, and scopes a PoC based on what actually exists. If the Blueprinting reveals AI is not viable at Kementan today, the report is still a valuable deliverable and Nera avoids a Rp 2B misallocation. If it validates, the estimated PoC+MVP value is Rp 2.5-3.5B.

Blueprinting Team

The Team and the Plan

Accounts utilize phased initialization with parallel scaling. Bireka's core architecture team tackles the hardest problems sequentially to ensure maximum quality, but hands off to a parallel integration team (Nera/KMJ) allowing deployment timelines to overlap safely without bottlenecking.

Bireka Delivery Team (Per Account)

Recommended Sequencing

Financial Consolidation

The consolidated investment required across the three priority accounts is summarized below. Manpower estimates reflect base delivery costs (COGS) structured under the sequential timeline framework.

Consolidated Project Pricing

Recurring Revenue Pipeline Post-MVP

All amounts in IDR. Base Delivery Costs (Nera COGS). Exchange rate: IDR 17,800/USD. Adjustments if rate moves beyond ±5%. Valid 90 days from document date.

Risk Assessment

PT Bhinneka Rekayasa Teknologi has analyzed the deployment risk landscape across all three target BUMN accounts. Key operational, technical, and regulatory risks, along with formal mitigation strategies, are outlined below.

Platform Advantage

Why Not Microsoft Copilot or IBM?

BUMNs will ask this question. The answer is that cloud vendors sell general-purpose tools that require internet connectivity, per-user seat licensing, and months of consulting to customize. NayaAI runs on-premise, works through WhatsApp, costs a fraction of seat-based pricing, and drives Nera hardware sales. The comparison is not close.

NayaAI runs on a 128GB unified-memory edge server deployed at the client's site. The database handles structured data, time-series telemetry, vector search, and geospatial queries in a single instance. A local AI model handles routine operations; a cloud AI service (Google Gemini) is used only for model improvement, carrying anonymized data. Each new client gets a domain plugin on a shared platform core, so each subsequent deployment is faster and cheaper. Proven in palm oil plantations (35,000+ ha), industrial mills (digital twin with 18-channel IoT at 6,300Hz), and environmental monitoring. Live demo: naya.bireka.id

Decisions Required

Immediate Next Steps

• Week 1: Budget allocation confirmed. Bireka begins team mobilization and platform preparation.
• Week 2: Joint kickoff with AirNav (or designated first account).
• Week 2: KMJ coordinates Ancol assessment scheduling.
• Week 3: Kementan stakeholder mapping and engagement planning begins.
• Ongoing: Bireka prepares client-facing presentation materials for live demos.

Growth Beyond MVP

Each MVP validates a product category that Nera can resell. AirNav from 3 sites to 15+. Ancol from 5 zones to 10, plus new revenue modules (F&B analytics, queue management). Kementan from 1 directorate to cross-ministry deployment. The adjacent market is larger: the AirNav blueprint applies to InJourney Airports. The Ancol blueprint applies to ITDC and TWC. Detailed market sizing and 5-year projections are provided in Appendix D of this proposal.

About Bireka & NayaAI

PT Bhinneka Rekayasa Teknologi (bireka.id) is an Indonesian technology company specializing in edge-native AI platforms for operationally complex environments. Bireka has developed NayaAI from the ground up to solve Indonesia's unique operational challenges: unreliable connectivity, informal communication cultures, multilingual and slang-heavy workforces, and the need for tamper-proof data integrity.

NayaAI has been proven in:

• Palm oil plantations (35,000+ hectares) — WhatsApp intelligence for field operations
• Palm oil mills (MillOS V2) — Digital twin SCADA with 18-channel IoT telemetry at 6,300Hz
• Environmental compliance (Cakrawala) — SPARING/SIMATAG/AQMS regulatory monitoring

The platform runs on a modern open-source stack (Next.js, PostgreSQL, Ollama) deployed on Apple Silicon hardware, achieving enterprise-grade performance at a fraction of traditional enterprise software costs. Live demo is available at naya.bireka.id.

Technical Platform Capabilities

NayaAI ships as a single deployable application with license-gated product contexts. The same platform core serves every domain (MillOS, Plantation Intelligence, Cakrawala, and new BUMN plugins). This core-and-plugin architecture ensures that each subsequent deployment is faster and cheaper, reusing database, AI, and auth systems.

Proven Core Capabilities

Deployment Specifications

Every NayaAI deployment runs edge-first with data sovereignty by design — all client data is processed and stored on-premise, avoiding public cloud dependencies for sensitive BUMN operations.

Edge Infrastructure Specifications

Market Opportunity Analysis

The three initial MVPs validate three product categories that Nera can resell across other state-owned enterprises (BUMN) and adjacent markets, creating a pipeline estimated at Rp 49B to Rp 70B over 5 years.

Adjacent BUMN Replication Targets

5-Year Revenue Projection (Rp)